I am a hobby translator on hiatus while I attended the Institute for Advanced Analytics. Prior to attending the Institute, I would translate Japanese pop-culture media from Japanese to English in my free time. I occasionally used AI translation tools, supplementing gaps or areas where the AI fell short with my cultural and historical knowledge of Japan. At the Institute, we have constant discussions on the ethics of AI. After our communications week— a week where we focused on data storytelling and data ethics—I wondered what the academic and professional view was on using AI in translation. Luckily, I have a cool and smart friend who I can just ask!
Emily Sedlacek earned a M.A. in English, a B.A. in English, a B.A. in French, and a minor in Communication from Appalachian State University. Her research focuses on the broader impacts of translation and rhetoric, and how they shape our understanding of gender representations in global literature. She taught Rhetoric and Composition while earning her M.A. at Appalachian State University. She currently works as an editor at E Source—a predictive data science, market intelligence, consulting, and advisory services organization.
Shannon: What are your initial thoughts on the integration of AI in literature translation? Do you view it as a positive development or are there concerns you have?
Emily: I think it’s a very complicated answer initially. I think, “wow this could be a useful tool;” however, there’s a lot to consider with translation for a couple of reasons. First, what are you trying to do? Are you trying to translate the essence of something or are you trying to translate a document word-for-word? It might be a pretty useful tool for word-for-word translation, but for idioms or ‘turn of phrase’ expressions, translation can get really complicated.
Second, AI is only as smart as the information you feed it. Again, my initial reaction is, “oh yeah, that seems like it could be really useful and really cool way to go about translation.” But at the same time, if I was using an AI translation tool, I’d want to know a few things. What dictionary was it fed with? What dictionaries were chosen? Is it just regular classroom French or is it teaching more up-to-date slang? These factors will impact the type and quality of translation you get.
Shannon: One of the concerns with AI translation is the potential loss of nuance and cultural context. How do you think AI tools handle the preservation of literary style, tone, and cultural subtleties in translated works?
Emily: I am a Debbie Downer because I say that they don’t. Even on the most basic level of “I put this sentence in English into Google Translate” and say, “translate this into French,” it’s going to have issues. There’s a certain amount of cultural knowledge that you need for a really solid translation. Gayatri Spivak talks about how you owe it to the language that you’re working with to be intimately familiar with the “mother tongue,” and it’s not just about the language itself. It’s about what’s happening around the language as well.
In France, L’Acadmie Francaise works to “preserve” the French language and it’s run by a bunch of French traditionalists. They’re known for being racist and sexist. So institutionalized French is very different from street French. For example, when you’re speaking street French, there are actually a lot of different ways to refer to people who might use gender-neutral pronouns outside of the masculine and feminine noun rules in French.
If you are only using a bot without that context, how is that bot going to navigate the politics of gender identity in France? You’re going to miss that in the translation. Anytime anyone translates something into French, they have to navigate this. Like, “oh crap, do I use the masculine, plural form of ‘they’ when I mean a group of 60 women with one man?” Because the bot would default to the masculine version in this example. However, that doesn’t represent the identities in the group. So, I think there is a lot to be lost when thinking of translation only as a science in which you put in an input, and you get out an output. In my opinion, there’s so much more to translation than this. I just don’t think the AI we have right now is quite there.
Shannon: In the long run, do you believe AI will significantly transform the way we approach literary translation, or do you see it more as a supplementary tool in the hands of human translators?
Emily: I absolutely do think that it will alter the way we translate because I think it’s going to alter the way we do everything. It’s already altering a lot of the ways we write. So as an editor, I’m having to adapt very quickly.
I do hope that as some of the newness wears off, we’ll be smarter about the way we use it. Not just “we” as in experts on certain topics but the general public. Because I do think it has a lot of potential to be a really cool tool for making things more accessible.
There will always be a surplus of content requiring translation, surpassing the capacity of available translators or publishing houses to keep up with the demand. Because like I said, translation is really hard. It’s an art and the people who do it are absolutely brilliant. Also, think about how many videos that are out there that people with hearing impairments don’t have access to because there’s not enough boots on the ground to translate it to closed captioning. So, in a way, it does make things way more accessible.
But again, there’s always danger to it. Thinking of it as this villain that’s going to take over the world isn’t accurate. However, I also don’t believe in asserting that this is the ultimate solution for the future. It’s not going to eliminate writers, translators, or editors. In a lot of ways, it’s very similar to when the Kindle came out. The media kept claiming that it was going kill the book industry. Here I am though; sitting around surrounded by books right now.
[Shannon: It’s true, there are a lot of books.]
AI is not going to kill the editing and translation industry. It’s just going to make the reading experience different for a lot of people.
It will take time to figure out the best way to use it and to teach people how to use it well. Is it a villain? No. Is it a hero? No. It is just a tool fundamentally, but one that needs to be assessed often for accuracy and bias.
Emily at the University of Warwick, UK – Women in World (-) Literature Conference.
[Interview was edited for clarity and brevity]
Columnist: Shannon Wells